From: Philippe Verdy (verdy_p@wanadoo.fr)
Date: Thu Jun 26 2003 - 15:37:05 EDT
On Thursday, June 26, 2003 8:16 PM, Elisha Berns <e.berns@computer.org> wrote:
> It would appear from your answer that even after implementing the
> algorithm to search the Unicode block coverage of a font, the actual
> comparison "data", that is which blocks to compare and how many code
> points, is totally undefined. Is there any kind of standard for
> defining what codepoints are required to write a given language? This
> seems like the issue that fontconfig gets around by using all those
> .orth files which define the codepoints for a given language. But is
> there any standardized set of language required codepoint definitions
> that could be used?
>
> Anyways, where is the up-to-date list of Unicode blocks to be found?
On the Unicode.org website or its published book.
> It's odd to think that the old way of using Charset identifiers in
> fonts worked a lot more cleanly for finding fonts matching a
> language/language group. I would think this kind of core issue would
> be addressed more cleanly by the font standard.
The ICU datafiles contain such list of codes needed to cover almost completely each combination of language+script.
Now these datafiles are shared across multiple implementations with the I18n initiative project, which tries to define a common source of locale data for multiple vendors (preiously this project was in li18nux.org now extended to cover other open systems than Linux, such as most BSD and Unix variants, with a joint effort with the GNU project and other Unix and Java solution providers)...
Of course, nothing forbids a particular text to use other characters than those strictly needed for a particular language...
This archive was generated by hypermail 2.1.5 : Thu Jun 26 2003 - 16:22:23 EDT